dispatching-parallel-agents

Warn

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: MEDIUMPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION] (MEDIUM): Indirect Prompt Injection Surface. The Agent Prompt Template recommended in the skill ingests potentially untrusted data ([error summary]) into a high-capability context (code modification). 1. Ingestion points: The [error summary] field in the dispatch template, which typically contains log data. 2. Boundary markers: Absent. The template does not use delimiters (e.g., XML tags or triple quotes) to isolate log data from task instructions. 3. Capability inventory: The workflow targets agents with write/execute permissions to fix source code. 4. Sanitization: No instructions provided for cleaning or validating external inputs. 5. Mitigation: The skill includes a manual integration and review step, which serves as a security control.
  • [NO_CODE] (SAFE): The skill is entirely markdown-based and does not include any scripts or executable files.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 16, 2026, 05:39 AM